Director of Offensive Security

The Company 

NorthMark Compute & Cloud (NMC²) is backed by dedicated leadership and investment, with a clear mission as it operates at the bleeding edge of technology. Its goal is to scale and enhance the high-performance computing (HPC) and cloud infrastructure that supports its clients’ research, production, and delivery, enabling breakthroughs that shape the industries of tomorrow. Its engineers build critical infrastructure to eliminate friction in scientific research, simulations, analysis, and decision-making, accelerating discovery and driving faster innovation. 

The Position 

The Director of Offensive Security reports directly to the CISO and owns continuous adversarial validation of the NMC² production environment. This is not a scheduled pentest function or a compliance-checkbox red team. You will build and run a standing offensive capability that operates against production with authorization, emulates named threat actors relevant to our customer base and infrastructure class, and produces independent, evidence-backed assessments of whether our controls work under realistic attack conditions. 

This function operates as an independent line of assurance within the Security organization, with a direct reporting relationship to the CISO. To preserve objectivity, assessment findings are delivered to the CISO without editorial review by the teams whose controls or systems are under evaluation. Security Engineering, Platform Engineering, and Security Architecture receive findings as remediation owners. 

Responsibilities: 

• Build and run a continuous red team program against the production NMC² environment: HPC clusters, multi-tenant Kubernetes, bare-metal provisioning infrastructure, customer network fabric, identity plane, and the internal control surface itself (SIEM, EDR, IAM, PAM) 

• Execute adversary emulation campaigns aligned to MITRE ATT&CK v15 TTPs relevant to our threat model: financially motivated access brokers (e.g., TTP sets associated with initial access brokers targeting financial services customers), APT groups with demonstrated interest in research computing and scientific workloads, and insider threat scenarios covering privileged operator abuse 

• Independently validate detection and response efficacy: every red team operation produces a detection coverage report measured against the SOC and IR functions, including time-to-detect, time-to-contain, and detection gap inventory by ATT&CK technique ID 

• Own the purple team feedback loop: every undetected TTP becomes a tracked detection engineering deliverable with owner and SLA, every detected-but-unresponded TTP becomes a tracked IR playbook deliverable 

• Run continuous attack surface validation against production, not just pre-production, with a documented rules-of-engagement framework, blast radius controls, and CISO-level authorization gates for destructive or high-risk techniques 

• Lead threat-led penetration testing of the HPC-specific attack surface: Slurm and workload manager abuse, GPU driver and firmware attack paths, InfiniBand and RDMA fabric isolation, scheduler privilege escalation, cross-tenant lateral movement in shared compute, and scientific software supply chain compromise 

• Own offensive validation of cloud and Kubernetes controls: IAM boundary testing, cross-account and cross-tenant escape attempts, container breakout chains, service mesh bypass, admission controller evasion, and secrets management integrity 

• Drive threat modeling at design stage for new platform capabiliti